Re: How do you use the loadings plot to find metabolites?
Reply #1 –
Greg,
So the real number that you refer to is the mass to charge ration (m/z) and the second number is the retention time. In your example the m/z is 313.3 and the retention time is 1313seconds. If you want to get a better resolution on the mass then you can do something like:
ix<-which(fill@groups[,"mzmed"] > 313.3 & fill@groups[,"mzmed"] < 313.4 & fill@groups[,"rtmed"] > 1313)
dim(fill@groups[ix,]) ## this will tell you how many results you have, hopefully only one :)
fill@groups[ix,] ## look at the results
However, from your question I'm not sure if you want to look at the loadings in more detail as well. You can access the loadings directly using something like:
loadings<-loadings(pca.result)
head(loadings)
The loadings are descriptors of the original variable. They tell you how the variables are connected. If you go back to the original data you may or may not see a statistical significant result with a t-test. I would suggest looking at some of the information on PCA. The simca website is very useful and holds a wealth of information.
Some good sites that I have found useful in the past:
http://www.umetrics.com/Content/Documen ... artI-3.pdf ## this is a nice introduction
http://www.metabolomics.se/Courses/MVA/ ... hu-Fri.pdf ## A lot of slides and progressive. A very good explanation, shows the reasons for scaling, which the example on the cookbook does not do.
http://www.statsoft.com/textbook/princi ... -analysis/